NVIDIA/mpi-acx

MPI accelerator-integrated communication extensions

CudaC++CMakefile
This is stars and forks stats for /NVIDIA/mpi-acx repository. As of 05 May, 2024 this repository has 29 stars and 4 forks.

MPI Accelerator Extensions Prototype This code provides a simple prototype for the proposed stream and graph triggered MPI Extensions, as well as kernel triggering for partitioned communication. The prototype currently supports a hybrid MPI+CUDA programming model. Requirements MPI-ACX requires CUDA 11.3 or later. The MPI library must support the partitioned communication API introduced in MPI 4.0. The MPI library must be initialized with support for the MPI_THREAD_MULTIPLE threading model. If communication...
Read on GithubGithub Stats Page
repotechsstarsweeklyforksweekly
hungps/flutter_pokedexDartC++CMake2.1k+20523+6
Gioni06/terminal.cssEJSCSSJavaScript9720520
asciinema/asciinema-serverElixirHTMLSCSS2.1k-12540
thehaigo/live_onnxElixirCSSHTML46050
atgreen/emacsEmacs LispTeXMakefile5000
snaiper80/hashids-erlangErlangMakefile200100
nitrogen/simple_bridgeErlangMakefileJavaScript1100750
g-andrade/stacktrace_compatErlangMakefile18050
chrislusf/seaweedfsGoJavaHTML18.4k02k0
kubernetes-sigs/metrics-serverGoMakefileShell5k+101.7k+3